Uncovering the community structure in signed social networks based on greedy optimization

The formality of signed relationships has been recently adopted in a lot of complicated systems. The relations among these entities are complicated and multifarious. We cannot indicate these relationships only by positive links, and signed networks have been becoming more and more universal in the study of social networks when community is being significant. In this paper, to identify communities in signed networks, we exploit a new greedy algorithm, taking signs and the density of these links into account. The main idea of the algorithm is the initial procedure of signed modularity and the corresponding update rules. Specially, we employ the Asymmetric and Constrained Belief Evolution procedure to evaluate the optimal number of communities. According to the experimental results, the algorithm is proved to be able to run well. More specifically, the proposed algorithm is very efficient for these networks with medium size, both dense and sparse.